Radiomics Techniques in Gastric Cancer Prognosis

Radiomics has emerged as a powerful, non-invasive approach to extract quantitative features from medical images and convert them into high-dimensional data that inform prognosis in gastric cancer. By analysing patterns of tumour texture, shape and spatial heterogeneity on computed tomography, multiphoton or other imaging modalities, radiomic pipelines can detect subtle variations invisible to the human eye. These data are then combined with advanced statistical methods or machine learning, including deep learning, to construct predictive models that estimate risks of lymph node involvement, peritoneal dissemination, treatment response and recurrence. Integration of radiomic signatures with clinical and pathological factors yields nomograms that support personalised treatment planning, guide surgical and chemotherapy decisions, and stratify patients into distinct prognostic groups. The capacity of radiomics to improve early detection of occult metastases and predict recurrence-free survival has global relevance, offering a scalable tool to enhance precision oncology in varied healthcare settings.

Research from Nature Portfolio

One study proposed a collagen signature derived from multiphoton analysis of the tumour microenvironment in the serosal layer of gastric lesions. Alterations in collagen features were quantified and combined with tumour size, differentiation status and nodal status in a competing-risk nomogram. This model demonstrated high discrimination and calibration for individual prediction of postoperative peritoneal metastasis in patients with serosal invasion. Another investigation developed a radiomic signature-based model using preoperative contrast-enhanced computed tomography to predict recurrence-free survival in locally advanced gastric cancer. Through penalised Cox regression, seven radiomic features were selected and validated in internal and external cohorts. The radiomic model achieved superior area under the receiver operating characteristic curve values compared with clinical staging alone, enabling more accurate identification of high-risk patients prior to surgery.

Radiomics Techniques in Gastric Cancer Prognosis publication trend

The graph below shows the total number of articles in radiomics techniques in gastric cancer prognosis across all publications each year (not limited to Nature Index journals).

Technical terms

Radiomics: The extraction and analysis of large numbers of quantitative features from medical images to characterise tumour properties.

Nomogram: A graphical calculator combining multiple predictors to estimate an individual’s probability of a clinical outcome.

Peritoneal metastasis: The spread of gastric cancer cells to the peritoneal surfaces, often undetectable by conventional imaging.

Texture analysis: Quantitative assessment of variations in pixel intensity and spatial relationships within an image region.

Deep learning: A subset of machine learning using neural networks with multiple layers to automatically learn hierarchical feature representations.

References

1. Predicting postoperative peritoneal metastasis in gastric cancer with serosal invasion using a collagen nomogram. Nature Communications (2021).
2. A radiomics-based model for predicting prognosis of locally advanced gastric cancer in the preoperative setting. Scientific Reports (2021).
3. Development and Validation of a Computed Tomography–Based Model for Noninvasive Prediction of the T Stage in Gastric Cancer: Multicenter Retrospective Study. Journal of Medical Internet Research (2024).
4. Development and validation of an individualized nomogram to identify occult peritoneal metastasis in patients with advanced gastric cancer. Annals of Oncology (2019).
5. Noninvasive Prediction of Occult Peritoneal Metastasis in Gastric Cancer Using Deep Learning. JAMA Network Open (2021).